planktonic community structure and carbon cycling in the ... · planktonic community structure and...

Planktonic community structure and carbon cycling in the

Arabian Sea as a result of monsoonal forcing:

the application of a generic model

J.C. Blackford *, P.H. Burkill1

Plymouth Marine Laboratory, Prospect Place, PL1 3DH, Plymouth, UK

Received 9 March 2001; accepted 20 June 2002

Abstract

The Arabian Sea exhibits a complex pattern of biogeochemical and ecological dynamics, which vary both seasonally and

spatially. These dynamics have been studied using a one-dimensional vertical hydrodynamic model coupled to a complex

ecosystem model, simulating the annual cycle at three contrasting stations. These stations are characterised by seasonally

upwelling, mixed-layer-deepening and a-seasonal oligotrophic conditions, respectively, and coincide with extensively measured

stations on the two JGOFS ARABESQUE cruises in 1994. The model reproduces many spatial and temporal trends in

production, biomass, physical and chemical properties, both qualitatively and quantitatively and so gives insight into the main

mechanisms responsible for the biogeochemical and ecological complexity. Monsoonal systems are typified by classical food

web dynamics, whilst intermonsoonal and oligotrophic systems are dominated by the microbial loop. The ecosystem model

(ERSEM), developed for temperate regions, is found to be applicable to the Arabian Sea system with little reparameterisation.

Differences in in-situ physical forcing are sufficient to recreate contrasting eutrophic and oligotrophic systems, although the lack

of lateral terms are probably the greatest source of error in the model. Physics, nutrients, light and grazing are all shown to play a

role in controlling production and community structure. Small-celled phytoplanktons are predicted to be dominant and sub-

surface chlorophyll maxima are robust centers of production during intermonsoon periods. Analysis of carbon fluxes indicate that

physically driven outgassing of CO2 predominates in monsoonal upwelling systems but ecological activity may significantly

moderate CO2 outgassing in the Arabian Sea interior.

D 2002 Elsevier Science B.V. All rights reserved.

Keywords: Ecological modeling; Arabian Sea; Carbon cycle; Plankton community; 1D water column model

1. Introduction

The Arabian Sea (Fig. 1) is an unusual and complex

ocean basin. Bounded to the north, west and east by

land and exceeding depths of 4000 m and although

tropical, the basin is continually influenced by the

seasonally reversing monsoonal winds (Swallow,

0924-7963/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.

PII: S0924 -7963 (02 )00182 -3

* Corresponding author. Tel.: +44-1752-633468; fax: +44-

1752-633101.

E-mail address: [email protected] (J.C. Blackford).1 Present address: Southampton Oceanography Centre, Uni-

versity of Southampton, Waterfront Campus, Southampton SO14

3ZH, UK.

www.elsevier.com/locate/jmarsys

Journal of Marine Systems 36 (2002) 239–267

1991). These winds produce complete reversal of the

surface currents and generate a range of contrasting

oceanographic conditions (Burkill et al., 1993a). Thus

surface conditions can range from eutrophy driven by

seasonal upwelling through to oligotrophy generated

by aseasonal strong stratification (Yoder et al., 1993).

Between June and September, the South West Mon-

soon (SWM) blows as the concentrated, low-level

Findlater Jet (Findlater, 1974). This creates strong

Ekman transport that induces upwelling of inorgani-

cally rich deep water in the north–west of the region

(Kindle, 2002), with downwelling and mixed layer

deepening observed to the south–east of the jet.

Although more moderate, the North–East Monsoon

(NEM, December–February) also entrains deep water

into the photic zone. By contrast, the inter-monsoon

periods, with weak and variable winds permit the

development of well-stratified systems largely devoid

of nutrients in the surface layer. In the south of the

region, away from the influence of the monsoon

winds, seasonal mixing is slight, with thermal strat-

ification and a deep nutricline predominating through-

out the year (Schott and McCreary, 2001).

SWM induced upwelling has been considered to

generate the highest rates of primary production while

the NEM has been thought to be biologically less

dynamic. Much of this understanding has arisen from

satellite derived, ocean colour images of the basin.

These have shown accumulation of high chlorophyll

concentrations at the end of the SWM with lower

concentrations at other times (Banse and English,

2000). For the main SWM period, satellite information

has been poor because of the dense monsoon cloud

base and the difficulties of working at sea during the

monsoon. To better understand the system particularly

at this time, international biogeochemical process

studies were planned as part of the Joint Global Ocean

Flux Study, and the results are now being published, for

example in Deep-Sea Research (Burkill et al., 1993a,b;

Van Weering et al., 1997; Smith, 1998, 1999, 2000,

2001; Burkill, 1999a; Gage et al., 2000; Pfannkuche

and Lochte, 2000) with a synthesis report produced by

Watts et al. (2002). Some of the results from these

studies are surprising. Although we now know that

primary production can reach 4 g C m� 2 day � 1 in the

vicinity of the Omani coast at the end of SWM

(Savidge and Gilpin, 1999), production during the

SWM can be lower than during the main NEM period

in the North East of the region (Marra et al., 1998;

Gunderson et al., 1998) where deep convective mixing

due to cold winds blowing off the Indian subcontinent

can generate production rates that average more than 1

g Cm� 2 day � 1 over the season (Wiggert et al., 2000).

Similarly experimental studies at sea have shown that

Fig. 1. Geographical position of the modelled stations in the Arabian Sea.

J.C. Blackford, P.H. Burkill / Journal of Marine Systems 36 (2002) 239–267240

phytoplankton growth rates were similar during both

monsoons (Landry et al., 1998).

During the stratified inter-monsoonal periods deep

chlorophyll maxima are observed at depths of 40–60

m. In the offshore, oligotrophic regions, deep chlor-

ophyll maxima are observed typically at around 80 m

(Savidge and Gilpin, 1999). In oligotrophic areas and

during the intermonsoon, the ecological community is

dominated by smaller cells (Jochem et al., 1993;

Burkill et al., 1993b) and recycling of carbon and

nutrients within the photic zone is also observed to be

an important feature. During the monsoons, large cells

such as diatoms and subsequently their meso-grazers

contribute more to cycling (Owens et al., 1993; Tarran

et al., 1999). Also of importance is the exudation of

dissolved organic carbon (DOC) and subsequent sec-

ondary production via bacterial utilisation, which,

although less variable, plays a significant part in

carbon cycling in the region (Pomroy and Joint,

1999; Ducklow, 1993).

The biogeochemical response to the seasonally

reversing monsoonal forcing is complex. Studies on

carbon dioxide in the Arabian Sea have shown a

substantial increase in dissolved inorganic carbon in

the coastal regions due to strong upwelling in the SW

monsoon. This was also accompanied by very high

CO2 partial pressures in surface waters (Mintrop et al.,

1999), which enhance the flux of carbon dioxide to the

atmosphere. High wind speeds also increase the rate of

exchange between atmosphere and ocean. The high

rates of primary production, and its subsequent respi-

ration or decomposition, may however influence CO2

levels in the photic zone. Physically mediated changes

in salinity and the biological influence on total alka-

linity via N speciation may also act to modify pCO2 in

surface waters. In well-mixed regions that experience

downwelling, the high levels of entrainment driven

primary production may lead to CO2 drawdown.

Where Ekman pumping is slight or absent and surface

primary production small, air–sea fluxes may be less

significant. Sinking of a significant proportion of

production is also observed (Buesseler et al., 1998),

seemingly associated with large cell communities in

the wake of monsoon events. Garrison et al. (2000)

states that considerable evidence links carbon cycling

and export to food web structure.

This study aims, by use of a coupled ecosystem-

physical model, to reproduce the ecological observa-

tions and quantify the biogeochemical fluxes of car-

bon in the Arabian Sea as they vary on both seasonal

and spatial scales. Given the complexity exhibited by

this system it is important to use an ecosystem model

Fig. 2. The ecosystem model flow diagram indicating the carbon and nutrient pathways between the functional groups.

J.C. Blackford, P.H. Burkill / Journal of Marine Systems 36 (2002) 239–267 241

capable of reproducing the variety of food web

structures found in the region. Hence the choice of

the European Regional Seas Ecosystem Model

(ERSEM, Baretta et al., 1995; Fig. 2) that simulates

the cycling of carbon, nitrogen, phosphorous and

silicon within size-resolved phytoplankton and zoo-

plankton communities and the microbial loop. In this

respect, the study extends on previous modelling of

the region that has employed simpler representations

of the ecosystem (e.g. McCreary et al., 1996; Keen et

al., 1997; Ryabchenko et al., 1998; McCreary et al.,

2001). The compromise employed here, in order to

keep the study computationally tractable, but also due

to the lack of boundary conditions is to confine the

study to a finally resolved 1D vertical physical model

and omit all horizontal processes with the exception of

upwelling. Although this lack of horizontal terms is a

significant omission form the model, studies have

shown that local forcing is the dominant mediator of

water column dynamics in the region during all but

the latter stages of the southwest monsoon (Fischer,

2000; Rochford et al., 2000).

A number of questions are also posed by this study.

Firstly, can the ecological model, developed for tem-

perate systems, reproduce the dynamics of the Arabian

Sea system? Secondly, to what extent can differences

in in-situ physical forcing generate the complex trends

in community structure across the region without

recourse to advective mechanisms? Thirdly, what

mechanisms cause specific community types to estab-

lish? Finally, does regional productivity influence the

Arabian Sea’s ability to be a source or sink of CO2?

The coupled model has been applied to three

contrasting stations in the Arabian Sea (Fig. 1, Table

1), which range from seasonally upwelling to near

a-seasonal oligotrophy. These stations were exten-

sively sampled by the two UK JGOFS ‘‘Arabesque’’

cruises held in 1994 as part of the international

JGOFS Arabian Sea Process Study (Burkill, 1999b).

The first cruise sampled the late SWM period, whilst

the second cruise sampled the autumn inter-monsoon

and early Northeast monsoon transition (INEM

cruise). Whilst the goal is to replicate and elucidate

the Arabesque cruises data set, reference is also made

to the 1994–1996 US JGOFS Arabian Sea expedition

(Smith et al., 1998a), which sampled a qualitatively

similar transect from the coast of Oman to the

Oceanic interior. Results from the US JGOFS and

ARABESQUE cruises are generally qualitatively and

quantitatively similar (Garrison et al., 2000).

The physical model system utilises the equations of

the Mellor–Yamada turbulence closure model and the

vertical diffusion submodel of the Princeton Ocean

Model (Blumberg and Mellor, 1980; Mellor and

Yamada, 1982). This coupled Princeton/Mellor–

Yamada/ERSEM model has been applied successfully

to water columns in the Adriatic (Allen et al., 1998;

Vichi et al., 1998a,b) and the Mediterranean Sea (Allen

et al., in press), but not previously to tropical regions.

2. The model

2.1. Physics and physical forcing

The Princeton/Mellor–Yamada physical model

determines vertical temperature, turbulent kinetic

energy and diffusion coefficient profiles. The model

is forced by sea surface temperature, salinity and wind

stress fields for 1994. Surface wind vectors are taken

from the ECMWF (European Centre for Medium-

range Weather Forecasts) model with a resolution of

12 h (Fig. 3). Daily values of sea surface temperature

are derived from the Comprehensive Ocean Atmos-

phere Data Set (COADS) using monthly mean values

averaged over the years 1993–1995. The inaccuracy

in modelled surface heat flux data (Weller et al., 1998)

and lack of a measured time series has precluded its

use here. Surface salinities are derived from the US

JGOFS database using qualitatively similar stations

(Table 2). Incident solar radiation is calculated as a

function of latitude and day length (Dobson and Smith,

1988), modified by monthly mean values of cloud

cover again derived from the COADS 1993–1995 data

set. Background light extinction coefficients have been

Table 1

Space– time coordinates of the modelled Arabesque cruise stations

Station Latitude

(jN)Longitude

(jE)Depth Arabesque cruise:

dates on station, 1994

C210 C212

A1 19 59 3397 4–7 Sept.,

30 Sept.

21–26 Nov.

A3 16 62 3927 13–14 Sept.,

28 Sept.

30–6 Nov/Dec.

A7 8 67 4705 19–21 Sept. 10–13 Dec.


chosen to fit the 1% light depths measured on station

(Table 2).

The physical model considers the top 200 m of the

water column divided into 40 equal layers of 5 m. The

lower boundary is considered to be adiabatic with the

exception of nutrients, CO2, O2, temperature and

salinity, which are held to constant values (Table 2).

At the upper boundary atmospheric exchange of CO2

and O2 are modelled. The partial pressure of CO2 in

the atmosphere is parameterised from Goyet et al.

(1998) in the range 335.0–355.0 Aatm with minimum

values recorded from October to December and max-

imum values recorded in January and February. Car-

bonate chemistry and the partial pressure of CO2 in the

water are calculated using the equations described in

Taylor et al. (1991), which utilise the Hansson (1973)

equilibrium coefficients and Weiss’ (1974) calculation

of CO2 solubility. In the absence of modelled pH,

salinity normalized total alkalinity is taken from Mill-

ero et al. (1998) (NTA= 2290 Amol kg� 1) and modi-

fied according to the surface salinity. Gas transfer

piston velocity is parameterised from Liss and Merli-

vat (1986), which differentiates between a smooth

surface regime (wind speed < 3.6 m s� 1), a rough

surface regime (3.6–13.0 ms� 1) and breaking waves

(>13.0 m s� 1). Oxygen saturation is calculated from

the formula of Weiss (1970) whilst the surface aeration

rate uses the empirical formula described in Allen et al.

(1998).

In order to reproduce the Ekman derived upwelling

observed in the region for station A1, an additional

vertical velocity has been parameterised from wind

speed and direction. The Princeton model derives

horizontal velocities for each layer (I) which may be

equated to a lateral loss of water mass (AI), a propor-

tion of which ( p) is replaced by water upwelling from

below, the remainder by lateral advective gains. So for

each layer, with UI representing upwelling loss and

UI � 1 upwelling gain:

UI�1 ¼ UI þ pAI

The constant p is chosen to give appropriate vertical

upwelling velocities, in this case 2 m day� 1 at 100 m

from Shi et al. (2000). Upwelling is assumed to occur

Fig. 3. Annual cycle of wind speed and direction for (a) station A1, (b) A3, (c) A7, as derived from the ECMWF model. Due North equates to

vertically upwards, West to the left, etc.


only when wind direction veers within 18j of true

south–west. This reproduces the observed seasonal

signal of upwelling in the region well (Halpern et al.,

1998; Rixen et al., 2000). The resulting upwelling

velocities for station A1 develop from May, peak in

July and subside by mid-September. The resulting

change in concentration for any state variable (CI) is

therefore given by:

DCI ¼ UI�1 � CI�1 � ðUI þ pAI Þ � CI

Thus simulations including upwelling are not conser-

vative and experience horizontal loss terms as well as

vertical gains. The simulation of downwelling is more

problematic as horizontal boundary conditions are

vital to formulating inputs to the system. With this

information unavailable a simulation of downwelling

has not been attempted.

2.2. Wind forcing

The ECMWF model provides a realistic spatial and

temporal distribution of wind vectors for the Arabian

Sea basin (Weller et al., 1998) (Fig. 3), which are in

close agreement with the winds measured during the

Arabesque cruises (Burkill, 1999b). These wind pat-

terns are the primary influence on the physical structure

of the water column, its seasonality and hence ecolog-

ical activity.

At station A1 (Fig. 3a), the (ECMWF model) SW

monsoon initially forms in May and following a

breakdown in early June becomes established from

mid-June onwards. Wind speeds exceed 13 m s� 1 for

the remainder of June, July and the first half of

August, thereafter dropping steadily, although the jet

maintains its direction until mid-September. The inter-

monsoon lasts only a month with the NE wind pattern

establishing itself by mid-October, breaking down in

early December and thereafter intermittent until the

end of January. Wind speeds are significantly less

when compared with the SW monsoon. At station A3

(Fig. 3b), the SW monsoon starts earlier, lasts longer

and has consistently higher wind speeds (>15 m s� 1)

than at A1. Additionally the NE monsoon has faster

mean wind speeds and is more consistent than at A1.

At station A7 (Fig. 3c), wind speeds are low and the

SW monsoon much less distinct than at either stations

A1 or A3.

2.3. The ecosystem model

The ecosystem model applied here is the ERSEM

model (Baretta et al. 1995, Fig. 2) which has been

applied in a variety of physical contexts from 1D to 3D

including the North Sea (Patsch and Radach, 1997;

Lenhart et al., 1997; Broekhuizen et al., 1995), the

Adriatic Sea (Allen et al., 1998; Vichi et al., 1998a,b)

and the Mediterranean Sea (Zavaterelli et al., 2000;

Allen et al., in press). For this study, the pelagic

components of ERSEM have been used. These and

the modifications specific to the Arabian Sea study are

briefly described. A mathematical description of

ERSEM can be found in the two ERSEM special

issues (Baretta et al., 1995; Ebenhoh et al., 1997 and

subsequent papers), the World Wide Web (http://

www.pml.ac.uk/ecomodels/ersem.htm) and particu-

larly in the individual papers cited below.

ERSEM is conceived as a generic model, which

when applied on a basin scale should be capable of

correctly simulating the spatial pattern of ecological

fluxes throughout the seasonal cycle and across

eutrophic to oligotrophic gradients (Baretta-Bekker

et al., 1997). The model uses three principle units

Table 2

Physical parameters and boundary conditions applied to each

modelled station

Unit Station

A1 A3 A7

Physical

Background

viscosity

Amol 1.0e� 5 1.0e� 5 1.0e� 5

Background

extinction

m� 1 0.04 0.04 0.04

Background silt mg�m� 3 750.0 350.0 250.0

Surface

pCO2 (air) A atm 350.0 350.0 350.0

Salinity max psu 36.40 36.50 36.65

Salinity min psu 35.70 36.00 35.25

Lower boundary

Temperature jC 15.0 15.0 14.0

Salinity psu 35.8 35.8 35.3

Oxygen mmol�m� 2 5.0 5.0 10.0

Carbon dioxide mmol�m� 2 2250.0 2250.0 2250.0

Phosphate mmol�m� 2 2.30 2.30 2.30

Nitrate mmol�m� 2 25.0 25.0 25.0

Ammonium mmol�m� 2 0.0 0.0 0.0

Silicate mmol�m� 2 30.0 30.0 30.0


of currency—carbon, nitrogen and phosphorus—with

each functional group containing these three pools.

The inclusion of the silicate cycle allows a distinction

between diatom and non-diatom production. Inor-

ganic nitrogen is sub-divided into ammonium and

nitrate compartments enabling the distinction between

new and recycled production.

Four functional groups describe the phytoplankton,

nominally picophytoplankton characterised as less

than 2 Am diameter, flagellates (between 2 and 20

Am), dinoflagellates (greater than 20 Am) and dia-

toms. Phytoplankton dynamics are mediated by pho-

tosynthesis (as a function of temperature and light

availability), respiration (both basal and active),

excretion (nutrient-stressed and active), lysis, mortal-

ity and predation. Phytoplankton contain internal

nutrient pools and thus varying C/N/P ratios. Nutrient

uptake is a function of the difference between the

internal and external pools (Ebenhoh et al., 1997).

Two size classes, the micro- and mesozooplankton

represent zooplankton. The processes of ingestion,

respiration, excretion and mortality contribute to the

overall growth rate of the population. Microzooplank-

ton, with variable internal C/N/P ratios (Baretta-Bek-

ker et al., 1995, 1997) predate on picophytoplankton,

flagellates and bacteria, whilst the mesozooplankton

compete for flagellates but also predate on the dia-

toms and dinoflagellates. Mesozooplankton are trea-

ted as a biomass based predation model, identical in

construct to the microzooplankton description. Given

the lack of diurnal processes and the lack of dynamic

mesozooplankton predators in the current application

the case for developing a vertical migration meso-

zooplankton model is slight. Further, given the con-

sistent warm temperatures in the region, it is

hypothesised that mesozooplankton are more respon-

sive than populations in cold temperate regions

(Roman et al., 2000) and so are more readily mod-

elled by a simple biomass approach.

Excretion and lysis products from all the ecolog-

ical functional groups appear as both dissolved and

particulate organic matter (DOM and POM), each of

which are here sub-divided into two classes. The

particulates are divided into large and small catego-

ries, the larger plankton (mesozooplankton, diatoms,

dinoflagellates, microzooplankton) contribute to the

large POM class with the smaller plankton producing

the smaller POM. The smaller POM is considered to

be more readably degradable by bacteria, due to a

higher surface area to volume ratio) and has a lower

sedimentation rate than its larger counterpart. The

dissolved fraction is divided into labile and semi-

labile fractions, with DOM production being divided

equally into each. This allows at least a crude repre-

sentation of the range of organic compounds in the

dissolved fraction, some of which may be directly

taken up by bacteria, some of which requires further

degradation before bacterial utilization. Parameters

relating to POM and DOM are included in Table 3.

Bacteria and heterotrophic flagellates represent the

microbial loop (Baretta-Bekker et al., 1995, 1997).

Bacteria utilise labile DOM and are also responsible

for the degradation of POM to DOM. Semi labile

DOM is considered to become labile at the rate of 1%

per day. Bacteria compete with phytoplankton for

nutrients and are predated by heterotrophic flagellates

Table 3

Bacteria model parameters

Description Unit Bacteria

Assimilation rate at 10 jC day� 1 8.38

Half saturation oxygen limitation – 0.3125

Fraction of small detritus

available for B1

– 0.05

Fraction of large detritus

available for B1

– 0.005

Assimilation efficiency – 0.4

Assimilation efficiency at low

oxygen concs

– 0.2

Basal respiration at 10 jC day� 1 0.3

Mortality rate day� 1 0.05

Max cell-quotum N mmol N

(mg C)� 1

0.0126

Max cell-quotum P mmol P

(mg C)� 1

0.000786

Michaelis constant for P uptake mmol P m� 3 0.5

Michaelis constant for N uptake mmol N m� 3 1.0

C/N ratio (Redfield) – 6.625

Dissolution of dissolved organics

to inorganics

day� 1 0.05

Relative nitrification rate day� 1 0.05

Sediment conc for 0.01 day� 1

nitrification rate

mg m� 3 2000.0

Rate of conversion of semi-labile

to labile DOM

day� 1 0.01

Sedimentation rate of large

POM class

m day� 1 7.5

Sedimentation rate of small

POM class

m day� 1 0.3


that are in turn grazed by the larger zooplankton. All

the consumers are also considered to graze within

their functional group and pseudo cannibalistic terms

are included. Fig. 2 summarises the interactions of the

functional groups. Ecosystem parameters are given in

Tables 3–7.

The simulations have been made using SESAME

(Software Environment for Simulation and Analysis of

Marine Ecosystems; Ruardij et al., 1995). The physical

model uses a semi-implicit finite difference scheme

forward in time and centered in space. The net trans-

port of the biological variables are calculated using the

physical model and then passed into SESAME where

they are integrated along with the biogeochemical rates

of change using a variable time step Euler method.

2.4. Initial conditions

Starting with vertical distributions of physics, chem-

istry and biology derived from literature and cruise

data, each modelled water column has been run for at

least 20 years using repeating forcing functions in order

to remove any perturbations due to initial conditions.

3. Model results

3.1. Physical variables

The physical model produces good approximations

to observed mixed layer depths (Figs. 4a and 5). The

annual cycle of thermal mixing and stratification

produced by the model (Fig. 6a) agrees well with

trends reported for the region (Angel, 1984; Gardner

et al., 1999). Stations A1 and A3 both exhibit strong

stratification during the inter-monsoon periods, with

pronounced mixing events associated with the SW

monsoon and to a lesser extent the NE monsoon.

During the monsoons, the base of the euphotic zone is

in the mixed layer, whilst during the inter-monsoon

periods it lies in the thermocline, consistent with

Brock et al. (1993). The SWM Arabesque cruise

coincides with a period of rapid decrease in simulated

mixed layer depth at both stations A1 and A3.

Relaxation after mixing events appears to be reason-

ably rapid. The INEM cruise coincides with the end

of a fairly stable simulated thermal structure and the

initial mixed layer deepening associated with the

Table 4

Optical phytoplankton parameters

Description Unit Value

Minimum value of optimal irradiance w m� 2 4.00

Maximum daily shift in optimal irradiance – 0.25

Photosynthetically available irradiance – 0.50

Adaptation depth m 10.00

Table 5

Phytoplankton model parameters

Description Unit Diatoms Flagellates Picoplankton Dinoflagellates

Assimilation rate (10 jC) day� 1 2.0 2.25 3.15 1.6

Basal respiration (10 jC) day� 1 0.2 0.2 0.2 0.2

Exudation under nut. stress – 0.05 0.10 0.15 0.05

Activity respiration – 0.2 0.45 0.5 0.45

Redfield N/C ratio mmol N (mg C)� 1 0.0126 0.0126 0.0126 0.0126

Minimal N/C ratio mmol N (mg C)� 1 0.00687 0.00687 0.00687 0.00687

Maximum N/C ratio mmol N (mg C)� 1 0.0252 0.0252 0.0252 0.0252

Redfield P/C ratio mmol P (mg C)� 1 0.786e� 3 0.786e� 3 0.786e� 3 0.786e� 3

Minimal P/C ratio mmol P (mg C)� 1 0.428e� 3 0.428e� 3 0.428e� 3 0.428e� 3

Maximum P/C ratio mmol P (mg C)� 1 0.157e� 2 0.157e� 2 0.157e� 2 0.157e� 2

Affinity for NO3 (mg C)� 1 day� 1 0.0025 0.0025 0.0025 0.0025

Affinity for NH4 (mg C)� 1 day� 1 0.01 0.01 0.015 0.01

Affinity for PO4 (mg C)� 1 day� 1 0.0025 0.0025 0.0025 0.0025

Nutrient stress threshold – 0.70 0.75 0.75 0.75

Nutrient stress sedimentation rate m day� 1 5.0 0.0 0.0 5.0

Minimal lysis rate day� 1 0.05 0.05 0.05 0.05

Si uptake Michaelis constant mmol Si m� 3 0.3 – – –

Standard Si/C ratio mmol Si (mg C)� 1 0.03 – – –


onset of the NEM. Annual cycles of mixed layer

depth and euphotic depth for A3 closely match the

mooring data for the similarly located US JGOFS

mooring (Kinkade et al., 2001). Station A7 exhibits a

deeper, essentially permanent thermal stratification in

the absence of strong wind forcing.

3.2. Nutrients

The modelled nitrate cycle (Fig. 6b) and modelled

profiles validated by measurements (Fig. 7) show a

progressive deepening of the nutricline from North to

South that is disturbed only by the monsoonal mixing

events. At A1 modelled surface concentrations of

nitrate exceed 5 mmol�m� 3 during the SW monsoon,

comparing well with observed values in the range 3–7

mmol�m� 3 (Woodward et al., 1999). The position of

the nutricline also agrees well with measurements (Fig.

7a). Modelled surface nitrate at A3 during the SWM

cruise period are lower than measured, ( < 0.5 mmol

m� 3 c/w >5 mmol m � 3). During the inter-monsoons

and at the permanently oligotrophic station A7, mod-

elled surface levels of nitrate are extremely low and

correspond very closely with measured values ( < 0.25

mmol m � 3 at A1, < 0.1 mmol�m � 3 at A3 and A7), as

do the nutricline depths as evidenced in Fig. 7d–f.

Both modelled and observed N/P ratios are low

( < 10.0) indicating phosphate limitation is not a fea-

ture of the region.

The range of ammonium as measured on the SWM

cruise is well reproduced by the model, concentra-

tions peak at 4.07 mmol�m� 3 at A1, 1.52 mmol m � 3

at A3 and 0.68 mmol m� 3 at A7, comparing with

measured peak values of 4.05, 1.91 and 0.74, respec-

tively (Woodward et al., 1999). In contrast with

measurements that show a more diffuse distribution

in the surface mixed layer, albeit highly variable,

modelled ammonium exhibits a sub-surface maxima

at all stations, deepening progressively with distance

from the Omani coast. Peak ammonium concentra-

tions are typically found above the nitrate nutricline

and just below the areas of peak phytoplankton

concentration.

Modelled silicate profiles show good agreement

with data at A1 and A7 with both surface mixed layer

concentrations and nutricline depth corresponding to

measurements (Woodward et al., 1999). A similar

mismatch as observed with the nitrate is evident at

A3, modelled surface layer silicate during the SWM

Table 6

Zooplankton model parameters

Description Unit Heterotrophic flagellates Microzooplankton Mesozooplankton

Assimilation rate at 10 jC day� 1 1.63 0.8 0.33

Food concentration where relative uptake is 0.5 mg C m� 3 120.0 80.0 40.0

Assimilation efficiency – 0.4 0.5 0.6

Fraction of excretion going to DOM – 0.5 0.5 0.5

Basal respiration rate at 10 jC day� 1 0.02 0.02 0.02

Oxygen saturation where respiration is 0.5 mmol m� 3 7.8125 7.8125 7.8125

Excreted fraction of uptake – 0.5 0.5 0.5

Mortality due to oxygen limitation day� 1 0.25 0.25 0.25

Temperature independent mortality day� 1 0.05 0.05 0.075

Max. quotum N/C mmol N (mg C)� 1 0.0167 0.0167

Max. quotum P/C mmol P (mg C)� 1 0.001 0.001

Lower threshold (mg C m� 3) for feeding mg C m� 3 15.0 10.0 1.0

Damping coefficient in N-excr. day� 1 0.5 0.5

Damping coefficient in P-excr. day� 1 0.5 0.5

Table 7

Feeding (preference) matrix

From To

Heterotrophs Microzoos Mesozoos

Bacteria 1.0 0.5 –

Picophytoplankton 1.0 1.0 0.1

Flagellates – 1.0 1.0

Diatoms – – 1.0

Dinoflagellates – – 1.0

Heterotrophic flagellates 0.2 1.0 0.5

Microzooplankton – 0.2 1.0

Mesozooplankton – – 0.2


being too low (0.4 c/w 1.7 mmol m� 3) and nutricline

depth too deep.

3.3. Primary producers

At each modelled station, inter-monsoonal periods

are characterised by deep production/biomass maxima

at 35 m (A1), 55 m (A3) and 75 m (A7), in reasonable

agreement with observed deep chlorophyll maxima at

35, 60 and 70 m, respectively (Barlow et al., 1999;

Savidge and Gilpin, 1999) (Fig. 8a). The biomass

associated with the deep maxima decline with distance

from the Omani coast, at the oligotrophic A7 being

about one third that of the biomass at A1. Modelled

monsoonal periods are characterised by surface

blooms at A1 and A3. In agreement with data the

Fig. 4. Comparisons of model results with data from the Arabesque cruises; (a) 1% light depth (m), (b) mixed layer depth (m), (c) phytoplankton

biomass (g C m� 2), (d) zooplankton biomass (g C m� 2), (e) bacterial biomass (g C m� 2), (f) primary production (g C m� 2 d� 1), (g) bacterial

production (g C m� 2 day� 1). Modelled values are given on the x-axis, observations on the y-axis. The solid line represents 1:1 correspondence.

The bars on (f) and (g) represent the range of measured data.


SW monsoonal bloom reaches deeper at A3 than at A1

in response to the deeper mixed layer and nutricline.

Depth integrated phytoplankton biomass reaches 6.6 g

C m � 2 at A1 and 6.2 g C m� 2 at A3 during the SW

monsoon. Significant blooms are also simulated dur-

ing the NE monsoon with peak phytoplankton biomass

at 3.9 and 3.3 g C m � 2 at A1 and A3, respectively.

The A7 simulation exhibits a very weak surface bloom

in response to mixing during the SWM period. Mod-

elled depth integrated biomass during the intermon-

soon periods show less of a spatial trend, the decline in

concentration from the Omani coast being offset by the

deepening of the euphotic zone. In comparison with

data, the model appears to over-predict phytoplankton

biomass at A7 during the INEM period, however the

general seasonal and spatial trends measured during

ARABESQUE are reproduced (Fig. 4c).

At A1 modelled primary production (Fig. 9) during

the SW monsoon agrees closely with that measured

during the ARABESQUE cruise (Savidge and Gilpin,

Fig. 5. Comparison of modelled (solid line) and measured (dots) temperature profiles for (a) station A1 during the SWM cruise; (b) A1, INEM

cruise; (c) A3, SWM; (d) A3, INEM; (e) A7, SWM; (f) A7, INEM.


Fig. 6. Annual cycle of (a) temperature (jC); (b) nitrate (mmol N m� 3); (c) Ammonium (mmol N m� 3) and (d) f-ratio for the top 150 m for

each station.


1999), but is slightly higher than that measured on

similar stations during the US JGOFS program (Barber

et al., 2001). Conversely modelled production at A1

during the remainder of the annual cycle agrees well

with Barber et al. (2001) but slightly over estimates

ARABESQUE measurements for the INEM period.

Modelled production at A3 is consistently lower than

both UK and US data, consistent with the under

estimation of nutrients in the model, although the

seasonal signal seems qualitatively correct.

Depth integrated f-ratios (Fig. 6d) are generally low

in all simulations, with annual means of 0.35, 0.24 and

0.24 for stations A1, A3 and A7, respectively. Season

minima are associated with inter-monsoon periods

(0.28, 0.11 and 0.13, respectively) and maxima with

the monsoon events (0.65, 0.89 and 0.41, respec-

tively). Locally, high f-ratios (>0.5) are associated with

the base of deep chlorophyll maxima at all stations,

where production is driven by the nutricline. Model

results agree well with measurements by Watts and

Fig. 7. Comparison of modelled (solid line) and measured (diamonds) nitrate profiles for (a) station A1 during the SWM cruise; (b) A1, INEM

cruise; (c) A3, SWM; (d) A3, INEM; (e) A7, SWM; (f) A7, INEM, (mmol N m� 3).


Fig. 8. Annual cycle of (a) primary production (mg C m� 3 day� 1); (b) bacterial biomass (mg C m� 3); (c) bacterial production (mg C m� 3

day� 1) and (d) total dissolved organic carbon (mg C m� 3), for the top 150 m for each station.


Owens (1999) who recorded f-ratios of 0.33, 0.24 and

0.21 (at stations A1, A3 and A7, respectively) during

the INEM cruise.

3.4. Bacteria and DOM

The model predicts that the semi-labile dissolved

pool, contributes 69%, 89% and 90% (for stations A1,

A3 and A7, respectively) to the total organic carbon

content, integrated over the 200 m depth range of the

simulations. The labile dissolved pool contributes 15%,

5% and 6%, respectively with particulates accounting

for 10%, 5% and 4%, respectively. However the

majority of labile DOC exists below the photic zone

(Fig. 8d), in areas of low oxygen concentration where

degradation by aerobic bacteria is minimal. In the well-

oxygenated mixed layer, concentrations of labile DOC

average about 40 mg Cm � 3, with peaks (60–80 mg C

m � 3) associated with the end of monsoon periods.

Labile DOC remains slightly elevated during the inter-

monsoon at A1 (50 mg C m� 3). The model predicts a

significant pool of semi-labile DOC which acts as a

carbon cycle buffer as hypothesized by Azam et al.

(1994) and reported by Ducklow et al. (2001), although

surface DOC in the model (labile + semi-labile) ranges

from 800 to 2300 mg C m � 3 compared with cruise

mean measurements of 888–984mg Cm� 3. Certainly

the basic assumption that half of produced DOC is

labile, half semi-labile is open to question, although the

lack of advective loss terms may also explain DOC

build up in the model water columns.

Measured total organic carbon (Hansell and Pelt-

zer, 1998), from the US JGOFS transect, show an

increase with distance from the Omani coast of about

1.0 mol C m� 2 and a seasonal decrease after the SW

monsoon of f 2.0 mol C m � 2. The model replicates

these trends, producing a mean increase of f 3.0 mol

C m � 2 with distance from Oman and a post mon-

soonal decrease of f 2.6 mol C m � 2. The discrep-

ancy in spatial trend may be due to the fact that the

US JGOFS transect did not reach as far into the

oligotrophic region as the Arabesque cruises. Both

model and data also show a vertical decrease in TOC,

although the modelled profiles (1.5–2.0 g C m � 3

surface, < 0.1 g C m � 3 at 150 m) are more exagger-

ated than the measured data (0.9 g C m� 3 surface,

0.5 g C m� 3 at 150 m).

Modelled estimates of bacterial biomass and pro-

duction are in reasonable agreement with data (Pomroy

and Joint, 1999; Wiebinga et al., 1997; Fig. 4e and g).

Modelled bacterial biomass shows a small increase

with distance from the Omani coast (0.7, 0.95 and 1.1

g Cm � 2 at A1, A3 andA7, respectively) not supported

by measurements, but show little of the seasonality and

spatial patterns exhibited by the autotrophs. Ducklow

et al. (2001) found that bacterial production was

suppressed during the NE monsoon and early SW

monsoon and slightly elevated during the spring

inter-monsoon, features that the model reproduces. A

maximum seasonal signal of f 1.7 fold increase over

the SW monsoon also agrees well with the maximum

< 2 fold increase observed by Ducklow et al. (2001).

Modelled net bacterial production is slightly higher at

A1 (206 mg C m� 2 day � 1) than at A3 or A7 (161 and

178 mg C m � 2 day � 1). It is notable that bacterial

production at A3 during the SWM cruise period is

consistently high in both model and data. In the model,

this is due to a short-lived but significant peak in DOC

production at the end of the monsoon period as

nutrient-stressed cells undergo lysis. This feature is

also simulated at A1, but does not coincide with the

cruise.

Fig. 9. Modelled annual cycles of primary production (solid line) for

(a) station A1 and (b) station A3, compared with data from the

ARABESQUE (w ) and US JGOFS (5) studies. US JOGFS station

S07 corresponds exactly with ARABESQUE station A3, The mean

of US JGOFS stations S02 and S03 has been used to compare with

ARABESQUE station A1.


3.5. Zooplankton

Modelled total zooplankton biomass (mesozoo-

plankton, microzooplankton and heterotrophic nano-

flagellates; Fig. 10) generally follows the pattern of

phytoplankton biomass, peaking 10–30 days after

monsoon phytoplankton blooms. The percentage of

zooplankton within total plankton decreases with dis-

tance from the Omani coast, being on average 43% at

A1, 32% at A3 and 26% at A7, as does the mean

zooplankton biomass, 2783, 1553 and 1151 mg C

m� 2, respectively (Table 8). During the later stages

of both monsoons, at A1 and A3, zooplankton stand-

ing stocks exceed that of phytoplankton. In contrast

with the picoplankton dominated phytoplankton, zoo-

plankton are dominated throughout the annual cycle

by the larger mesozooplankton class at station A1.

This supports observations that the microbial loop is

an important source of food for zooplankton during the

inter-monsoon (Madhupratap and Gopalakrishnan,

1996). At A3 the zooplankton are a more mixed

community and whilst mesozooplankton are still the

largest group by biomass, both heterotrophic flagel-

lates and microzooplankton are significant. At A7

Fig. 10. Depth integrated annual cycles for the zooplankton community: (a) station A1, (b) A3, (c) A7, (mg C m� 2).


heterotrophic flagellates join mesozooplankton as the

dominant grazers by biomass. The model reproduces

the measured spatial and seasonal trends (Stelfox et al.,

1999) although zooplankton biomass is consistently

over estimated, in comparison with the ARABESQUE

data, by a factor of 2.85F 0.5. However other surveys

(e.g. Sorokin et al., 1985; Reckermann, 1996; Smith et

al., 1998b; Roman et al., 2000) have recorded higher

zooplankton standing stocks in the region that corre-

spond closely with modelled values. For example

Roman et al. (2000) measured an inshore–offshore

biomass range of f 1.2 to 0.5 mg C m � 2 during the

inter-monsoon and f 3.2 to 0.75 during the late SW

monsoon. The range of model results is respectively

1.1–0.5 and 3.3–0.9 mg C m� 2.

Modelled bactivory is higher in upwelling systems

throughout the year, showing sustained levels around

150 mg C m � 2 day � 1 during the intermonsoon and a

decline during the monsoon periods. Short-lived peaks

in bactivory (>300 mg C m� 2 day � 1) are predicted at

the end of bothmonsoons. Bactivory at the oligotrophic

station is constant at around 100 mg C m � 2 day � 1.

Bactivory by heterotrophic nanoflagellates dominates

that of microzooplankton throughout the annual cycles.

These results are consistent with observations of hourly

bacterial grazing rates (Weisse, 1999).

Estimates of mesozooplankton herbivory derived

from the ARABESQUE transect are very variable,

ranging from 3–64 mg C m� 2 day � 1 at A7 during

the intermonsoon to 24–204 mg C m� 2 day � 1 at A1

during the SWM cruise. Modelled mesozooplankton

herbivory shows good agreement to the spatial and

seasonal trend measured, with herbivory of f 100 mg

C m � 2 day � 1 at A7 during the intermonsoon and

f 680 mg C m � 2 day � 1 at A1 during the SW

monsoon. Applying the same factor derived from the

differences between measured and modelled zooplank-

ton biomass, modelled herbivory is consistent with

estimates of herbivory from the field. Mesozooplank-

ton ingestion measured on the US JGOFS transect

(Roman et al., 2000), whilst also variable, compare

well with those modelled (Table 9). Microzooplankton

herbivory of 29–496 mg C m� 2 day� 1 accord well

with measurements by Edwards et al. (1999) of 161–

300 mg C m� 2 day � 1. Estimations of microzoo-

plankton herbivory in the north east of the region

during the NE monsoon (Reckermann and Veldhuis,

1997) of 38% of total phytoplankton biomass, 67% of

total production, 49% of picophytoplankton biomass

and 102% of picophytoplankton production compare

with model results of 28%, 67%, 42% and 100%,

respectively.

3.6. Modelled plankton community structure

From the perspective of annual means (Table 8),

within the trend of decreased plankton biomass with

distance from the Omani coast, the contribution of

phytoplankton to total biomass remains constant at just

under 50%. Zooplankton, from been almost on a par

with phytoplankton at A1 (43%) decreases to 26% of

the total at A7, whilst bacteria shows the reverse trend,

increasing from 11% at A1 to 26% at A7. This

compares with ranges of 43–64% for phytoplankton

and 16–44% for bacteria recorded on the US JGOFS

transect (Garrison et al., 2000).

Table 8

Annual mean biomass and community structure for each modelled

station

Station A1 A3 A7

Total biomass (mg C m� 2) 6521 4889 4348

% Contribution to Phytoplankton 46 49 48

biomass total Zooplankton 43 32 26

Bacteria 11 19 26

% Contribution to Diatoms 22 9 7

phytoplankton total Flagellates 16 22 23

Picoplankton 58 60 53

Dinoflagellates 4 9 18

% Contribution to

zooplankton total

Heterotrophic

flagellates

27 30 37

Microzooplankton 15 24 15

Mesozooplankton 58 46 48

Table 9

Mesozooplankton ingestion rates (mg C m� 2 day� 1) from the

model and the US JGOFS process study

Period S02 Data A1 Model S07 Data A3 Model

Late NE monsoon 390 431 139 154

Spring inter-monsoon 456 318 702 98

Late SW monsoon 527 628 349 344

Early NE monsoon 708 371 606 110

Modelled station A3 is compared with the co-incident US station

S07, modeled station A1 is compared with the nearby US station

S02. The Hirst –Sheader estimates from Roman et al. (2000) are

used.


The composition of the phytoplankton community

shows distinct spatial and temporal trends (Fig. 11).

Generally picophytoplankton are the dominant func-

tional group, in agreement with observations (Tarran

et al., 1999), contributing 50–60% of inter-monsoonal

phytoplankton biomass and up to 90% during the

monsoonal community succession. Diatoms, flagel-

lates and picophytoplankton are associated with mon-

soonal forcing whilst the dinoflagellates are generally

absent from monsoonal periods. Diatoms only contrib-

ute significantly during periods of monsoonal forcing

when they can account for over 70% of phytoplankton

biomass, again reflecting observations, e.g. Garrison

et al. (2000) and Tarran et al. (1999).

At station A1 the four phytoplankton classes

exhibit contrasting behaviour throughout the annual

cycle (Fig. 11a). Competition between diatoms and

dinoflagellates, which otherwise have a similar model

niche depends on the availability of silicate. Hence

diatoms are predicted only in a narrow band in the

vicinity of the nutricline during inter-monsoon periods

and distributed throughout the mixed layer during the

monsoon seasons, the f-ratio for diatoms being con-

sistently greater than 0.4. The distribution of the

dinoflagellates is restricted to the remaining parts of

the photic zone where they are predominantly depend-

ent on recycled nutrients ( f-ratiof 0.2). Picophyto-

plankton and flagellates show similar distributions

during intermonsoon periods, associating with both

the nutricline and nutrient recycling ( f-ratios = 0.2–

0.4). For all functional groups, competition and sea-

sonal succession is enhanced during the monsoon

bloom period with diatoms and picophytoplankton

showing oscillating negatively correlated distribu-

tions. Towards the end of the SW monsoon, when

the mesozooplankton grazers are dominant, carnivory

of the micro-grazers provides a niche for a flagellate

bloom.

Although the gross community structure at A3 is

similar to that at A1, certain differences are evident. At

A3 (Fig. 11b), diatoms are much more restricted than

at A1 occurring only during the late SW monsoon ( f-

ratios>0.4). Dinoflagellates replace diatoms for all but

the SW monsoon period. Flagellates exploit the late

development of diatoms, showing higher monsoonal

densities than at A1.

Two distinct communities appear to exist at A7

(Fig. 11c), a mixed layer community of flagellates and

dinoflagellates deriving nutrients from recycling ( f-

ratiosf 0.2) and a diatom ( f-ratiof 0.5) and pico-

phytoplankton community ( f-ratiof 0.3) associated

with the sub-surface maxima and dependant more on

new production. This accords well with observations

by Jochem et al. (1993).

The annual mean ratio of bacterial production (BP)

to primary production (PP) increases with distance

from the Omani Coast (0.17 at A1, 0.38 at A3 and

0.49 at A7). Plots of bacterial production against

primary production for the spring inter-monsoon,

SW and NE monsoon periods (Fig. 12a) show a

number of distinct regimes, replicated by similar

plots of bacterial biomass to phytoplankton biomass

(Fig. 12b). During inter-monsoon periods at A3 and

A7 bacterial and primary production appear to be

strongly coupled and BP/PP ratios are about 0.45.

This direct coupling is not evident during the inter-

monsoon at A1, when the BP/PP ratio is f 0.2.

Although the BP/PP ratio during monsoon periods at

A1 and A3 exhibits similar values (f 0.1), A3 tends

to exhibit a smoother relationship, whilst at A1 the

relationship is far more chaotic. At A7 during the

SW monsoon, although the BP/PP ratio remains at

similar levels to the intermonsoon period, the direct

coupling has ceased and perturbation is evident.

Analysing biomass to production ratios for both

bacteria and phytoplankton indicates that A1 is

characterized by high ratios throughout the annual

cycle (f 0.43 for phytoplankton) whilst the ratio at

A3 and A7 is consistently lower (f 0.17). Hence

grazing is far more important at A1 (grazing

amounting to 20–40% per day) than at the other

stations (grazingV 15% per day) and is evidenced

by a more chaotic relationship between bacteria and

phytoplankton. On the other hand, the monsoonal

periods are characterized by mixing induced pertur-

bations, shifting the size structure of the phytoplank-

ton community and the response time of the grazing

community. The modelled NE monsoon is seen to

be far less dynamic, exhibiting relatively little per-

turbation in bacterial or phytoplanktonic response.

During the intermonsoon periods at A3 and A7, the

coupling between bacterial and phytoplankton pop-

ulations is initially due to bacterial dependence on

DOC production by phytoplankton and latterly direct

bacterial competition for nutrients with the phyto-

plankton.


Fig. 11. Annual cycles of the phytoplankton functional groups at each station (a) diatoms, (b) flagellates, (c) picoplankton, (d) dinoflagellates

(mg C m� 3).


3.7. Fluxes

Table 10 gives the modeled annual mean carbon

fluxes between key functional groups for each station.

Net CO2 uptake by the primary producers shows little

spatial variation between each station with fluxes

through picophytoplankton dominating, in agreement

with Savidge and Gilpin (1999) and Barber et al.

(2001). A shift away from diatoms and towards

smaller size classes with distance from the Omani

coast is also produced by the model agreeing with

observations by Latasa and Bidigare (1998) and Bar-

low et al. (1999). The fate of assimilated CO2 does

however show marked spatial variation with POC and

DOC production, generally via nutrient stress induced

lysis, accounting for 58% of net production at A1 but

84% at A3 and 92% at A7. In contrast grazing is

significantly greater at A1 than A3 and, in turn, A7,

with herbivory accounting for 39%, 15% and 8% of

net CO2 uptake, respectively, and carnivory account-

ing for 38%, 31% and 25% of secondary production,

respectively.

Table 11 illustrates the same carbon fluxes for

different seasons and monsoonal sub-seasons at station

A1. Primary production by the small picophytoplank-

ton class dominates throughout. Diatom production

peaks are associated with both the NEM and SWM.

Diatom lysis reaches a maximum in early August.

Peak herbivory and carnivory is associated with the

SW monsoon. Grazing is the most variable component

Fig. 12. (a) Bacterial against primary production (mg C m� 2 day� 1) and (b) bacterial against phytoplankton biomass (mg C m� 2) for three

contrasting periods. The spring intermonsoon (Julian days 90–150), denoted by I, the SW monsoon (Julian days 200–240), denoted by S and

the NE monsoon (Julian days 15–45), denoted by N for each modelled station, (A)1, (A)3 and (A)7. The NE monsoon for A7 is not shown as it

coincides with the inter-monsoon at A7.


of the system, whilst the microbial loop and the

associated remineralisation the least variable.

Significant annual plankton community fixation of

CO2 is only evident at A1 (216 g C m � 2 year� 1),

photosynthesis and respiration at stations A1 and A3

are more balanced (26 and 14 g C m� 2 year � 1 net

fixation respectively). Despite the higher fixation,

significant CO2 out-gassing is indicated at station A1

Table 11

Seasonal carbon fluxes at station A1 (mg C m� 2 day� 1)

NEM SIM SWM AIM

345-

70

71-

150

151-

210

210-

270

271-

345

Net phytoplankton CO2 uptake

by Picophytoplankton 1845 2294 2193 2339 1939

by Flagellates 264 403 376 452 456

by Diatoms 285 87 971 561 124

by Dinoflagellates 11 112 59 96 45

Total net CO2 uptake 2405 2896 3599 3448 2564

Excretion/lysis to DOC/POC

by Picophytoplankton 967 1739 1273 1190 1315

by Flagellates 112 275 186 162 259

by Diatoms 156 50 401 284 60

by Dinoflagellates 5 50 30 3 15

by Heterotrophic flagellates 252 161 287 418 228

by Microzooplankton 132 111 100 93 73

by Mesozooplankton 322 242 563 586 310

Total phytoplankton loss 1240 2114 1890 1639 1649

Total consumer loss 706 514 950 1097 611

by Bacteria 41 50 43 39 42

Bacterial production

Bacteria 1802 2231 2024 1895 1910

Herbivory




Total herbivory 1107 725 1512 1668 879

Bactivory

of Bacteria 131 161 174 174 161

Carnivory




Total carnivory 420 303 696 620 360

Respiration

by Bacteria 1585 1966 1751 641 1665

by grazers 570 411 704 815 481

System budget

Sedimentation 150 152 158 341 212

Loss of CO2 to Atmospheric 9 � 12 193 213 � 27

Net community

CO2 uptake

250 517 1144 904 417

Table 10

Annual mean carbon fluxes (mg C m� 2 day� 1), for each modelled

station

A1 A3 A7

Net phytoplankton CO2 uptake

by Picophytoplankton 2106 1783 1932

by Flagellates 385 402 544

by Diatoms 366 83 62

by Dinoflagellates 49 103 213

Total net CO2 uptake 2906 2371 2751

Excretion/lysis to DOC/POC

by Picophytoplankton 1304 1522 1802

by Flagellates 200 327 495

by Diatoms 174 64 53

by Dinoflagellates 23 83 181

by Heterotrophic flagellates 259 96 77

by Microzooplankton 103 58 23

by Mesozooplankton 387 100 69

Total phytoplankton loss 1701 1996 2531

Total consumer loss 749 254 169

by Bacteria 43 58 73

Bacterial production

Bacteria 1974 2291 2804

Herbivory




Total herbivory 1132 359 211

Bactivory

of Bacteria 159 103 106

Carnivory




Total carnivory 428 110 53

Respiration

by Bacteria 1725 2072 2553

by grazers 578 225 158

System budget

Sedimentation 198 54 36

Loss of CO2 to Atmospheric 62 � 13 � 8

Net community CO2 uptake 603 74 40


during the SWmonsoon (Fig. 13) when daily loss rates

of CO2 can exceed 0.5 g C m� 2 day � 1. At this time-

modelled values of pCO2 agree well with measure-

ments of Goyet et al. (1998). Annually a flux of 22 g C

m� 2 to the atmosphere is predicted, whilst small draw

downs of CO2 at A3 (5 g C m � 2) and A7 (3 g C m� 2)

are simulated, in contrast to the generally predicted net

regional outgassing throughout the region (Goyet et

al., 1998; Kortzinger et al., 1997).

Sedimentation loss, measured at 200 m lags the

peak in production by a few weeks (Fig. 13), and over

the annual cycle is amounts to f 15% of net produc-

tion at each station. Of total community CO2 uptake,

sedimentation accounts for 33% at A1 and 90% at A3

and A7. At the latter two stations, the remaining

community uptake is accounted for by a build up of

organic carbon, which in reality may be advected

away. At A1, horizontal advection loss generated by

the modelled upwelling accounts for the difference

between community uptake and sedimentation. Plau-

sibly the mass advected out of the water column would

be subject to sedimentation downstream. Sedimenta-

tion rates are seen to be most closely associated with

larger cell production or the classical foodweb (dia-

toms and mesozooplankton), with sedimentation

events lagging 15–30 days behind production peaks,

agreeing with the analysis of Rixen et al. (2000). A

maximum simulated rate of f 400 mg C m � 2 day� 1

or approximately 20% of primary production at A1

agrees very well with measurements of f 300 and

17–28% for the late South West Monsoon (Buesseler

et al., 1998). Honjo et al. (1999) observed a double

maxima in sedimentation patterns during both the late

South West and North East Monsoons, which is

reproduced for the SWM in the model. The first

maximum is associated with the diatom/meso-grazers

Fig. 13. Annual cycles of sedimentation loss (mg C m� 2 day� 1) and CO2 flux to the atmosphere (mg C m� 2 day� 1) for (a) station A1, (b) A3

and (c) A7.


peak around Julian day 230, whilst the second is a

product of a subsequent flagellate/meso-grazer peak at

day 280. Modelled sedimentation rates for the NE

Monsoon are in the order of half the magnitude of SW

monsoon events, again in accordance with data (Honjo

et al., 1999). Monsoon sedimentation peaks are pre-

dicted to consist predominantly of mesozooplankton

fecal pellets, rather than ungrazed diatoms. Microbial

food webs, dominant during the inter-monsoon periods

and at the permanently oligotrophic stations, contrib-

ute little to sedimentation loss, averaging about 25 mg

C m � 2 day � 1, agreeing well with measurements of

22 mg C m � 2 day � 1 (Pollehne et al., 1993).

Grazer-autotroph oscillations are clearly seen in the

net community CO2 uptake, in association with the

latter part of the monsoonal periods. All simulations

indicate periods of net community CO2 production

that may indicate periods of grazer control.

4. Discussion

Despite the one-dimensional physical aspects of the

coupled model, the simulations generate a high pro-

portion of observed features in the region. A clear

distinction between monsoonal and inter-monsoonal

systems is shown and further distinctions between the

North East and South West monsoons are clear. Trends

of increasing oligotrophy away from the Omani coast

are also reproduced. State variables and process rates

generally validate well. Thus, if a reasonable approx-

imation to upwelling is parameterised, then the state

and rate variable dynamics are reproducible by con-

sidering water column processes only. Given that the

simulations differ only in wind forcing, cloud cover,

irradiance, background sediment loads, surface salinity

and lower boundary temperature then one can infer

that the local physical forcing alone can explain much

of the complex spatial trends found in the region, in

accordance with Fischer (2000).

The biggest errors in the model simulations occur at

station A3. This could be a simple error in timing in the

model, however the peak modelled surface nitrate

during the SWM at A3 is only 1.7 mmol m � 3

compared with observations of the order of 5 mmol

m � 3, suggesting that a more fundamental process is

missing from the A3 simulation. This is also evidenced

by the profile mismatch for SWM nitrate at A3 (Fig.

7c–d) which shows the modelled nutricline to be too

diffuse and too deep. It seems likely that the model’s

failure to inject sufficient nutrients into the photic zone

is the root cause of the production shortfall at A3.

Although not in the area generally thought to experi-

ence upwelling, a possibility is that the station was

affected by an upwelling eddy, the US studies have

shown filaments in this region (Dickey et al., 1998).

Potentially the simulation of entrainment driven mix-

ing is not sufficient to inject the observed quantities of

nutrients to the surface layer. Lee et al. (2000) found

that entrainment is at least seasonally important. One

conclusion is that resolution to diurnal radiation forc-

ing would be a valuable addition to the model

(McCreary et al., 2001).

The model supports the findings of Savidge and

Gilpin (1999) and Jochem et al. (1993) demonstrating

the potential for significant primary production in the

oligotrophic regions of the Arabian Sea. Under-esti-

mates of primary production in the model may be due

to the relatively high parameterisation of respiration

used, but the modelled production underpins the

dynamics of the microbial loop, which shows good

agreements with data. Evidence of higher than

assumed respiration rates in the region is given by

Robinson and Williams (1999). Furthermore Dickson

et al. (2001) found that stations with higher net

production to gross production ratios also had higher

DOC production rates compared to community respi-

ration, a feature that the model reproduces. Presumably

regions that respire less of the gross production pro-

duce cells with higher C/N ratios that are more prone to

nutrient stress lysis and increased DOC production.

This is certainly an important dynamic in the ERSEM

model as a consequence of the Droop kinetics used to

simulate nutrient limitation in phytoplankton. Lysis

induced by low internal cell nutrient quotas can

produce cell mortality rates of 10–50% day � 1. How-

ever model simulations using external nutrient con-

centrations to limit production show serious inconsis-

tencies with measured data. Better estimates of

respiration and DOC production rates are undoubtedly

required.

Model results suggest that up to 90% of organic

carbon in the upper Arabian Sea exist as DOC. The

principle source of DOC is identified as nutrient stress

lysis of phytoplankton. Two DOC maxima exist, one

in surface waters, the other, significantly larger, below


the photic zone, impinging on the oxygen minimum

zone characteristic of some Arabian Sea regions. This

build up of sub-surface DOC in the model is due to

mixing down and subsequent lack of advective lateral

transport and possibly anaerobic decomposers. How-

ever, it is indicative of an important sink of organic

carbon and concurs with observed accumulations of

DOC in productive systems (Copin-Montegut and

Avril, 1993; Carlson et al., 1994; Williams, 1995).

Degradation of surface DOC in the model is restricted

because bacterial growth rate is limited by competition

for nutrients and to some extent bacterial biomass is

controlled by predators, as demonstrated in a model-

ling context by Thingstad et al. (1997). Unlike most

other system components, which decline towards the

central Arabian Sea, neither bacteria nor mixed layer

DOC show a pronounced spatial trend. Whilst bacteria

become more nutrient constrained in oligotrophic

waters, increased respiration, with increased temper-

ature, offsets internal nutrient depletion and grazing

pressure declines. Accumulation of DOC in surface

layers may act to moderate the response of bacteria to

temporal trends in production (Azam et al., 1994). This

combination of factors may account for constant

bacterial populations.

The model supports findings that the non-monsoo-

nal Arabian Sea can generally be characterised as one

in which small celled phytoplankton communities

dominate (Tarran et al., 1999; Jochem et al., 1993;

Burkill et al., 1993b). The model suggests that 80–

95% of production is converted to DOC via cell lysis,

providing a substrate for the microbial loop. Without

perturbation, as evidenced by relatively constant bac-

terial production to primary production ratios, a multi-

vorous grazing community develops, heterotrophic

nanoflagellates grazing on picophytoplankton and

bacteria with mesozooplankton grazing the larger

phytoplankton cells and other zooplankton. The bulk

of these dynamics are associated with sub-surface

maxima with a combination of light and nutrients

controlling the system.

The two monsoonal events that affect the North

West of the region provoke complex spatial-temporal

changes in the modelled planktonic community struc-

ture. The initial response to the SW monsoon at A1

are concurrent blooms of diatoms, flagellates and

picophytoplankton driven by new production. The

subsequent development of an omnivorous grazing

community acts to suppress flagellates and causes

predator prey oscillations alternately favouring dia-

toms and picophytoplankton. Towards the end of the

monsoon, grazing pressure on diatoms and picophy-

toplankton creates a niche for flagellates that exploit

the remaining now mainly recycled nutrients in the

surface mixed layer. Similar patterns are modelled

during the NE monsoon at A1 and both monsoons at

A3, although blooms associated with the NE monsoon

are dominated by the < 2.0 and < 20.0 phytoplankton

size classes. The weaker mixing during the NE mon-

soon tends to favour the non-siliceous phytoplankton

groups. This is consistent with theories that silica

availability and hence diatom growth in the region is

largely driven by strong coastal upwelling as a con-

sequence of SW monsoonal forcing (Young and Kin-

dle, 1994). Low f-ratios throughout the region indicate

that the remaining phytoplankton communities are

frequently dependent on recycled nutrients. Thus

where upwelling is weak or non-existent diatoms are

out competed. The model reproduces the strong

observed coupling between monsoons, classical food

webs and sedimentation (Garrison et al., 2000).

The model system used does not include a precise

formulation of CO2 partial pressure and subsequent

air–sea CO2 exchange. Salinity forcing and subse-

quent estimates of total alkalinity are crude, as are

estimates of atmospheric partial pressure. However,

despite its limitations a qualitative estimation of air–

sea CO2 exchange is appropriate. Comparison of

modelled pCO2 in the water with data from Goyet et

al. (1998) for the same region, indicate that peak

modeled values are in the middle of the recorded range

during the monsoon period at A1. This gives credence

to the models prediction of out-gassing although

model estimates are generally significantly higher.

Conversely the model suggests draw down at A7 in

contrast to outgassing derived by Goyet et al. (1998).

Both modelled and measured pCO2 at A7 show no

seasonal signal, although model estimates are slightly

lower. Despite inaccuracies, the model confirms the

physically driven nature of CO2 flux into the atmos-

phere in regions of upwelling, but suggests biologi-

cally mediated draw down may be a significant factor

in determining the CO2 flux in regions of downwelling

during monsoonally mixed conditions.

There are challenges in validating models of this

kind against spatially and temporally specific data


generated on cruise based surveys. As shown by both

data and model, the Arabian Sea exhibits considerable

heterogeneity in both time and space. This is partic-

ularly true comparing the model output with the data

generated during the September Arabesque cruise.

This is a transitional period when the SW monsoon

breaks down and changes of up to an order of

magnitude of both biomass and fluxes occur within a

short period of time. A second challenge is the provi-

sion of forcing functions for the physical model. The

day-to-day variability in wind fields is crucial in

providing a reasonable degree of mixing in the model

and wind patterns are required to determine the spatial

trends in the region. It is thus important to use realistic,

high-resolution wind data. Although in this case,

winds measured during the survey coincide well with

the wind forcing used, the actual wind history is

unknown. As the behaviour of the system at any given

time is complexly linked to the detailed history of wind

over the preceding few days, weeks and months, the

use of modelled wind forcing must be seen as a

possible source of short term error in the model.

5. Conclusions

This modelling exercise illustrates that an ecosys-

tem model developed within the context of temperate

shelf sea issues can, without fundamental changes,

successfully simulate the mixed layer dynamics of

tropical ocean systems and further the dynamics of

an extremely heterogeneous system like the Arabian

Sea. A key factor is the provision of an accurate

physical environment and a model with sufficient

complexity to represent microbial, multivorous and

classical communities or the continuum of trophic

pathways as described by Legendre and Rassoulzade-

gan (1995). As the ecological model process descrip-

tions and parameters remain constant throughout this

study, the results show complex and contrasting com-

munities and exchanges may result from differences in

a few basic physical parameters and forcing fields.

Although the differential wind forcing across the

region and the proximity of the nutricline to the surface

at the onset of mixing events primarily determine the

spatial-temporal signal in productivity across the

region, the model shows that sub surface chlorophyll

maxima are also important and robust areas of pro-

duction agreeing with Gunderson et al. (1998). The

contrasting distributions of nutrients, silicate driven by

physical processes, nitrate by a combination of physics

and nitrification, ammonium via recycling are shown

to be more important in determining large cell com-

munity structure. Ecological niches in the latter part of

monsoon seasons are created by differential grazing,

both herbivorous and carnivorous. Regional trends in

microbial community structure are primary driven by

the fate of primary production, generally the propor-

tion undergoing short-term lysis. Ecological activity is

also suggested to be a significant moderator of CO2 out

gassing in the Arabian Sea interior.

By capturing much of the system’s features in a

simple 1D approach, the model underlines the useful-

ness of both fixed and lagrangian long-term water

column studies. These provide an important mecha-

nism for the assembly of detailed high frequency

process orientated data sets in remote locations, which

are crucial to understand the more complex aspects of

the system. Indeed simulations of dynamic monsoo-

nally driven events indicate that sampling programs

may be severely biased by their precise timing, whilst

programs in oligotrophic regions may be more oppor-

tunistic in their execution.

Although successful, the model is limited by the

lack of horizontal advective processes and resulting

turbulent kinetic energy terms. It may also be

improved by resolving diurnal physical processes

(McCreary et al., 2001). Many developments of the

ecosystem model could also be beneficial, for example

including coccolithophores, further resolving the phy-

toplankton or including behavioral and developmental

processes for mesozooplankton. The coupling of the

ecological model to a fully 3D representation of the

Arabian Sea would provide a further test of the

ecosystem model and subsequently an extremely

powerful tool for quantifying and exploring the Ara-

bian Sea system’s dynamics.

Acknowledgements

This research forms part of the Microbially Driven

Biogeochemistry programme of the Plymouth Marine

Laboratory. It was part funded by Defence Evaluation

Research Agency Grant (TQ/10/3/2) as well as the

UK Natural Environment Research Council and forms


part of the UK contribution to the SCOR/IGBP

JGOFS Arabian Sea Process Study. The authors

would also like to acknowledge the constructive and

helpful comments made by the anonymous referees.

References

Allen, J.I., Blackford, J.C., Radford, P.J., 1998. An 1-D vertically

resolved modelling study of the ecosystem dynamics of the

middle and southern Adriatic Sea. Journal of Marine Systems

18, 265–286.

Allen, J.I., Sommerfield, P.J., Siddorn, J., 2001. Primary and bacte-

rial production in the Mediterranean Sea: a modelling study.

Journal of Marine Systems (in press).

Angel, M.V., 1984. Marine science of the north-west Indian Ocean

and adjacent waters. Deep Sea Research I 31, 573–1035.

Azam, F., Steward, G.F., Smith, D.C., Duklow, H.W., 1994. In:

Lal, D. (Ed.), Significance of Bacteria in Carbon Fluxes in

the Arabian Sea. Indian Academy of Science, Bangalore, pp.

243–253.

Banse, K., English, D.C., 2000. Geographical differences in season-

ality of CZCS derived phytoplankton pigment in the Arabian

Sea for 1978–1986. Deep Sea Research II 47, 1623–1677.

Barber, R.T., Marra, J., Bidigare, R.C., Codispoti, L.A., Halpern,

D., Johnson, Z., Latasa, M., Goericke, R., Smith, S.L., 2001.

Primary productivity and its regulation in the Arabian Sea during

1995. Deep-Sea Research II 48, 1127–1172.

Baretta, J.W., Ebenhoh, W., Ruardij, P., 1995. The European region-

al seas ecosystem model, a complex marine ecosystem model.

Netherlands Journal of Sea Research 33 (3/4), 233–246.

Baretta-Bekker, J., Baretta, J.W., Rasmussen, E., 1995. The micro-

bial food web in the the European regional seas ecosystem

model. Netherlands Journal of Sea Research 33, 363–379.

Baretta-Bekker, J.G., Baretta, J.W., Ebenhoh, W., 1997. Microbial

dynamics in the marine ecosystem model ERSEM II with de-

coupled carbon assimilation and nutrient uptake. Journal of Sea

Research 38, 195–212.

Barlow, R.G., Mantoura, R.F.C., Cummings, D.G., 1999. Monsoo-

nal influence on the distribution of phytoplankton pigments in

the Arabian Sea. Deep-Sea Research II 46, 677–700.

Blumberg, A.F., Mellor, G.L., 1980. A coastal ocean numerical

model. In: Sunderman, J., Holtz, K.P. (Eds.), Mathematical

Modelling of Estuarine Physics, Proceedings of the Internation-

al Symposium, Hamburg, Aug. 1978. Springer-Verlag, Berlin,

pp. 203–214.

Brock, J., Sathyendranath, S., Platt, T., 1993. Modelling the season-

ality of subsurface light and primary production in the Arabian

Sea. Marine Ecology Progress Series 101, 209–221.

Broekhuizen, N.R., Heath, M.R., Hay, S.J., Gurney, W.S.C.,

1995. Modelling the dynamics of the North Seas mesozoo-

plankton. Netherlands Journal of Sea Research 33 (3/4),

381–815.

Buesseler, K., Ball, L., Andrews, J., Benitez-Nelson, C., Belastock,

R., Chai, F., Chao, Y., 1998. Upper ocean export of particulate

organic carbon in the Arabian Sea derived from thorium-234.

Deep-Sea Research II 46, 2461–2488.

Burkill, P.H., 1999a. ARABESQUE: UK JGOFS process studies in

the Arabian Sea. Deep Sea Research II 46, 529–863.

Burkill, P.H., 1999b. ARABESQUE: an overview. Deep-Sea Re-

search II 46, 529–548.

Burkill, P.H., Mantoura, R.F.C., Owens, N.J.P., 1993a. Biogeo-

chemical cycling in the northwestern Indian Ocean: a brief over-

view. Deep-Sea Research II 40, 643–650.

Burkill, P.H., Leakey, R.J.G., Owens, N.J.P., Mantoura, R.F.C.,

1993b. Synechococcus and its importance to the microbial food-

web of the northwestern Indian Ocean. Deep-Sea Research II

40, 773–782.

Carlson, C.A., Ducklow, H.W., Michaels, A.F., 1994. Annual flux

of dissolved organic carbon from the euphotic zone in the north-

western Sargasso Sea. Nature 371, 405–408.

Copin-Montegut, G., Avril, B., 1993. Vertical distribution and tem-

poral variation of dissolved organic carbon in the north–west

Mediterranean Sea. Deep-Sea Research I 40, 1963–1972.

Dickey, T., Marra, J., Sigurdson, D.E., Weller, R.A., Kinkade, C.S.,

Zedler, S.E., Wiggert, J.D., Langdon, C., 1998. Seasonal var-

iability of bio-optical and physical properties in the Arabian

Sea: October 1994–October 1995. Deep-Sea Research II 45,

2001–2026.

Dickson, M.-L., Orchardo, J., Barber, R.T., Marra, J., McCarthy,

J.J., Sambrotto, R.N., 2001. Production and respiration rates in

the Arabian Sea during the 1995 Northeast and Southwest Mon-

soons. Deep-Sea Research II 48, 1199–1230.

Dobson, F.W., Smith, S.P., 1988. Bulk models of solar radiation at

sea. Quarterly Journal of the Royal Meteorological Society 114,

165–182.

Ducklow, H.W., 1993. Bacterioplankton distributions and produc-

tion in the northwestern Indian Ocean and Gulf of Oman, Sep-

tember 1986. Deep-Sea Research II 40, 753–772.

Ducklow, H.W., Smith, D.C., Campbell, L., Landry, M.R., Quinby,

H.L., Steward, G.F., Azam, F., 2001. Heterotrophic bacterio-

plankton in the Arabian Sea: Basinwide response to year-

round high primary productivity. Deep-Sea Research II 48,

1303–1324.

Ebenhoh, W., Baretta, J.W., Baretta-Bekker, J.G., 1997. The pri-

mary production model in a marine ecosystem model ERSEM

II. Journal of Sea Research 38, 169–172.

Edwards, E.S., Burkill, P.H., Stelfox, C.E., 1999. Zooplankton her-

bivory in the Arabian Sea during and after the SW monsoon,

1994. Deep-Sea Research II 46, 843–863.

Findlater, J., 1974. The low-level cross equatorial air current of the

western Indian Ocean during the northern summer. Weather 29,

411–416.

Fischer, A.S., 2000. Ph.D. Dissertation, Massachusetts Institute of

Technology-Woods Hole Oceanographic Institute, 222.

Gage, J.D., Levin, L.A., Wolff, G.A., 2000. Benthic processes in the

Deep Arabian Sea. Deep Sea Research II 47, 1–7.

Gardner, W.D., Gundersen, J.S., Richardson, M.J., Walsh, I.D.,

1999. The role of seasonal and diel changes in mixed-layer

depth on carbon and chlorophyll distributions in the Arabian

Sea. Deep-Sea Research II 46, 1833–1858.

Garrison, D.L., Gowing, M.M., Hughes, M.P., Campbell, L., Caron,


D.A., Dennett, M.R., Shalapyonok, A., Olson, R.J., Landry,

M.R., Brown, S.L., Liu, H.-B., Azam, F., Steward, G.F.,

Ducklow, H.W., Smith, D.C., 2000. Microbial food web struc-

ture in the Arabian Sea: a US JGOFS study. Deep-Sea Research

II 47, 1387–1422.

Goyet, C., Millero, F.J., O’Sullivan, D.W., Eischeid, G., McCue,

S.J., Bellerby, R.G.J., 1998. Temporal variations of pCO2 in

surface seawater of the Arabian Sea in 1995. Deep-Sea Research

I 45 (4–5), 609–624.

Gunderson, J.S., Gardner, W.D., Richardson, M.J., Walsh, I.D.,

1998. Effects of monsoons on the seasonal and spatial distribu-

tions of POC and chlorophyll in the Arabian Sea. Deep-Sea

Research II 45, 2103–2132.

Halpern, D., Freilich, M.H., Weller, R.A., 1998. Arabian Sea sur-

face winds and ocean transports determined from ERS-1 scat-

terometer. Journal of Geophysical Research-Oceans 103,

7799–7806.

Hansell, D.A., Peltzer, E.T., 1998. Spatial and temporal variations of

total organic carbon in the Arabian Sea. Deep-Sea Research II

45, 2171–2194.

Hansson, I., 1973. A new set of pH scales and standard buffers for

sea water. Deep-Sea Research I 20, 479–491.

Honjo, S., Dymond, J., Prell, W., Ittekkot, V., 1999. Monsoon con-

trolled export fluxes to the interior of the Arabian Sea. Deep-Sea

Research II 46, 1859–1902.

Jochem, F.J., Pollehne, F., Zeitzschel, B., 1993. Productivity regime

and phytoplankton size structure in the Arabian Sea. Deep-Sea

Research II 40, 711–736.

Keen, T.R., Kindle, J.C., Young, D.K., 1997. The interaction of

southwest monsoon upwelling, advection and primary produc-

tion in the northwest Arabian Sea. Journal of Marine Systems

13, 61–82.

Kindle, J.C., 2002. The surface circulation of the northern Arabian

Sea. In: Watts, L., Burkill, P.H., Smith, S.L. (Eds.). Report of the

Indian Ocean Synthesis Group on the Arabian Sea Process

Study. JGOFS Report 35, 1–114.

Kinkade, C.S., Marra, J., Dicky, T.D., Weller, R., 2001. An annual

cycle of phytoplankton biomass in the Arabian Sea, 1994–

1995, as determined by moored optical sensors.

Kortzinger, A., Duinker, J.C., Mintrop, L., 1997. Strong CO2 emis-

sions from the Arabian Sea during the south–west monsoon.

Geophysical Research Letters 24, 1763–1766.

Landry, M.R., Brown, S.L., Campbell, L., Constantinou, J., Liu,

H.B., 1998. Spatial patterns in phytoplankton growth and micro-

zooplankton grazing in the Arabian Sea during monsoon forc-

ing. Deep-Sea Research II 45, 2353–2368.

Latasa, M., Bidigare, R.R., 1998. A comparison of phytoplankton

populations of the Arabian Sea during the spring intermonsoon

and Southwest monsoon of 1995 as described by HPLC-ana-

lyzed pigments. Deep-Sea Research II 45, 2133–2170.

Lee, C.M., Jones, B.H., Brink, K.H., Fischer, A.S., 2000. The

upper-ocean response to monsoonal forcing in the Arabian

Sea: seasonal and spatial variability. Deep-Sea Research II 47,

1177–1226.

Legendre, L., Rassoulzadegan, F., 1995. Plankton and nutrient dy-

namics in marine waters. Ophelia 41, 153–172.

Lenhart, H.J., Radach, G., Ruardij, P., 1997. The effects of river

input on the ecosystem dynamics in the continental coast zone

of the North Sea using ERSEM. Journal of Sea Research 38,

249–274.

Liss, P.S., Merlivat, L., 1986. Air– sea gas exchange rates: intro-

duction and synthesis. In: Buat-Menard, P. (Ed.), The role of

Air –Sea Exchange in Geochemical Cycling. D. Reidel, Dor-

drecht, pp. 113–127.

Madhupratap, M., Gopalakrishnan, T.C., 1996. Lack of seasonal

and geographic variation in mesozooplankton biomass in the

Arabian Sea and its structure in the mixed layer. Current Science

71, 863–868.

Marra, J., Dickey, T.D., Ho, C., Kinkade, C.S., Sigurdson, D.E.,

Weller, R.A., Barber, R.T., 1998. Variability in primary produc-

tion as observed from moored sensors in the central Arabian Sea

in 1995. Deep-Sea Research I 45, 2253–2268.

McCreary, J.P., Kohler, K.E., Hood, R.R., Olson, D.B., 1996. A

four component ecosystem model of biological activity in the

Arabian Sea. Progress in Oceanography 37, 193–240.

McCreary, J.P., Kohler, K.E., Hood, R., Smith, S., Kindle, J.,

Fischer, A., Weller, R.A., 2001. Influences of diurnal and intra-

seasonal forcing on mixed-layer and biological variability in

the central Arabian Sea. Journal of Geophysical Research 106,

7139–7155.

Mellor, G.L., Yamada, T., 1982. Development of a turbulence clo-

sure model for geophysical fluid problems. Reviews of Geo-

physics and Space Physics 20 (4), 851–875.

Millero, F.J., Lee, K., Roche, M., 1998. Distribution of alkalinity in

the surface waters of the major oceans. Marine Chemistry 60,

111–130.

Mintrop, L., Kortzinger, A., Duinker, J.C., 1999. The carbon diox-

ide system in the northwestern Indian Ocean during the south–

west monsoon. Marine Chemistry 64, 315–336.

Owens, N.J.P., Burkill, P.H., Mantoura, R.F.C., Woodward, E.M.S.,

Bellan, I.E., Aiken, J., Howland, R.J.M., Llewellyn, C.A., 1993.

Size-fractionated primary production and nitrogen assimilation

in the northwestern Indian Ocean. Deep-Sea Research II 40,

697–710.

Patsch, J., Radach, G., 1997. Long term simulation of the eutrophi-

cation of the North Sea: temporal development of nutrients,

chlorophyll and primary production in comparison to observa-

tions. Journal of Sea Research 38, 275–310.

Pfannkuche, O., Lochte, K., 2000. The Biogeochemistry of the

Deep Arabian Sea. Deep Sea Research II 47, 2615–2628.

Pollehne, F., Klein, B., Zeitzschel, B., 1993. Low light adaptation

and export production in the deep chlorophyll maximum layer in

the northern Indian Ocean. Deep-Sea Research II 40, 737–752.

Pomroy, A., Joint, I., 1999. Bacterioplankton activity in the surface

waters of the Arabian Sea during and after the 1994 SW mon-

soon. Deep-Sea Research II 46, 767–794.

Reckermann, M., 1996. Ultraphytoplankton and protozoan com-

munities and their interactions in different marine pelagic eco-

systems (Arabian Sea and Baltic Sea). Meereswissenschaftliche

Berichte Warnemunde 14, 139 pp.

Reckermann, M., Veldhuis, J.W., 1997. Trophic interactions be-

tween picophytoplankton and micro- and nanozooplankton in

the western Arabian Sea during the NE monsoon 1993. Aquatic

Microbial Ecology 12, 263–273.


Rixen, T., Haake, B., Ittekkot, V., 2000. Sedimentation in the west-

ern Arabian Sea the role of coastal and open-ocean upwelling.


Robinson, C., Williams, P.J.LeB., 1999. Plankton net community

production and dark respiration in the Arabian Sea during Sep-

tember 1994. Deep-Sea Research II 46, 745–766.

Rochford, P.A., Kindle, J.C., Gallacher, P.A., 2000. Sensitivity of

the Arabian Sea mixed layer to 1994–1995 operational wind

products. Journal of Geophysical Research 105, 14141–14162.

Roman, M., Smith, S., Wishner, K., Zhang, X., Gowing, M., 2000.

Mesozooplankton production and grazing in the Arabian Sea.


Ruardij, P., Baretta, J.W., Baretta-Bekker, J.G., 1995. SESAME: a

software environment for simulaton and analysis of marine

ecosystems. Netherlands Journal of Sea Research 33 (3/4),

261–270.

Ryabchenko, V.A., Gorchakov, V.A., Fasham, M.J.R., 1998. Sea-

sonal dynamics and biological productivity in the Arabian Sea

euphotic zone as simulated by a three-dimensional model. Glob-

al Biogeochemical Cycles 12, 501–530.

Savidge, G., Gilpin, L., 1999. Seasonal influences on size frac-

tionated chlorophyll a concentrations and primary production

in the north–west Indian Ocean. Deep-Sea Research II 46,

701–724.

Schott, F.J., McCreary, J.P., 2001. The monsoon circulation of the

Indian Ocean. Progress in Oceanography 51, 1–123.

Shi, W., Morrison, J.M., Bohm, E., Manghnani, V., 2000. The

Oman upwelling zone during 1993, 1994 and 1995. Deep-Sea

Research II 47, 1227–1248.

Smith, S.L., 1998. The 1994–1996 Arabian Sea expedition: oceanic

response to monsoonal forcing, part 1. Deep-Sea Research II 45,

1905–2501.



1532–1964.



1177–1677.



1069–1402.

Smith, S.L., Codispoti, L.A., Morrison, J.M., Barber, R.T., 1998a.

The 1994–1996 Arabian Sea Expedition: an integrated, inter-

disciplinary investigation of the response of the northwestern

Indian Ocean to monsoonal forcing. Deep-Sea Research II 45,

1905–1915.

Smith, S., Roman, M., Prusova, I., Wishner, K., Gowing, M.,

Codispoti, L.A., Barber, R., Marra, J., Flagg, C., 1998b. Sea-

sonal response of zooplankton to monsoonal reversals in the

Arabian Sea. Deep-Sea Research II 40, 2369–2404.

Sorokin, Y.I., Kopylov, A.I., Mamaeva, N.V., 1985. Abundance and

dynamics of microzooplanktonin the central tropical Indian

Ocean. Marine Ecology Progress Series 24, 27–41.

Stelfox, C.E., Burkill, P.H., Edwards, E.S., Harris, R.P., Sleigh,

M.A., 1999. The structure of zooplankton communities, in the

2 to 2000 Am size range, in the Arabian Sea during and after the

SW monsoon, 1994. Deep-Sea Research II 46, 815–842.

Swallow, J., 1991. Circulation in the northwestern Indian Ocean. In:

Smith, et al. (Eds.). US JGOFS Arabian Sea Process Study

Planning Report 13, pp. 37–39.

Tarran, G.A., Burkill, P.H., Edwards, E.S., Woodward, E.M.S.,

1999. Phytoplankton community structure in the Arabian Sea

during and after the SW monsoon, 1994. Deep-Sea Research II

46, 655–676.

Taylor, A.H., Watson, A.J., Ainsworth, A., Robertson, J.E., Turner,

D.R., 1991. A modelling investigation of the role of phytoplank-

ton in the balance of carbon at the surface of the North Atlantic.

Global Biogeochemical Cycles 5 (2), 151–171.

Thingstad, T.F., Hagstrom, A., Rassoulzadegan, F., 1997. Accumu-

lation of degradable DOC in surface waters: is it caused by a

malfunctioning of the microbial loop? Limnology and Ocean-

ography 42, 398–404.

Van Weering, T.C.E., Helder, W., Schalk, P., 1997. Netherlands

Indian Ocean program (1992–1993) first results. Deep Sea Re-

search II 44, 1177–1193.

Vichi, M., Pinardi, N., Zavatarelli, M., Matteucci, G., Marcaccio,

M., Bergamini, M.C., Frascari, F., 1998a. One dimensional

ecosystem model results in the Po prodelta area (northern

Adriatic Sea). Environmental Modelling and Software 13,

471–481.

Vichi, M., Zavatarelli, M., Pinardi, N., 1998b. Seasonal modulation

of microbially mediated carbon fluxes in the northern Adriatic

Sea—a model study. Fisheries. Oceanology. Journal of Marine

Systems 7, 182–190.

Watts, L.J., Owens, N.J.P., 1999. Nitrogen assimilation and the

f-ratio in the north-western Indian Ocean during an intermon-

soon period. Deep-Sea Research II 46, 725–744.

Watts, L., Burkill, P.H., Smith, S.L., 2002. Report of the Indian

Ocean Synthesis Group on the Arabian Sea Process Study.

JGOFS Report 35, 1–114.

Weiss, F., 1970. The solubility of nitrogen, oxygen and argon in

water and seawater. Deep Sea Research 17, 721–735.

Weiss, R.F., 1974. Carbon dioxide in water and sea water: the

solubility of a non-ideal gas. Marine Chemistry 2, 203–215.

Weisse, T., 1999. Bactivory in the north-western Indian Ocean dur-

ing the intermonsoon-north –east monsoon period. Deep-Sea

Research II 46, 795–814.

Weller, R.A., Baumgartner, M.F., Josey, S.A., Fischer, A.S., Kindle,

J.C., 1998. Atmospheric forcing in the Arabian Sea during

1994–1995: observations and comparisons with climatology

and models. Deep-Sea Research II 45, 1961–2000.

Wiebinga, C.J., Veldhuis,M.J.W., DeBaar, H.J.W., 1997. Abundance

and productivity of bacterioplankton in relation to seasonal up-

welling in the northwest Indian Ocean. Deep-Sea Research I 44,

451–476.

Wiggert, J.D., Jones, B.H., Dickey, T.D., Brink, K.H., Weller, R.A.,

Marra, J., Codispoti, L.A., 2000. The Northeast Monsoon’s im-

pact on mixing, phytoplankton biomass and nutrient cycling in

the Arabian Sea. Deep-Sea Research II 47, 1353–1386.

Williams, P.J.LeB., 1995. Evidence for seasonal accumulation of

carbon-rich dissolved organic material, its scale in comparison

with changes in particulate material and the consequential

effect on net C/N assimilation ratios. Marine Chemistry 51,

17–29.


Woodward, E.M.S., Rees, A.P., Stephens, J.A., 1999. The influence

of the SW monsoon upon the nutrient biogeochemistry of the

Arabian Sea. Deep-Sea Research II 46, 571–592.

Yoder, J.A., McClain, C.R., Feldman, G.C., Esaias, W.E., 1993.

Annual cycles of phytoplankton chlorophyll concentrations in

the global ocean: a satellite view. Global Biogeochemical

Cycles 7, 181–193.

Young, D.K., Kindle, J.C., 1994. Physical processes affecting avail-

ability of dissolved silicate for diatom production in the Arabian

Sea. Journal of Geophysical Research 99, 22619–22632.

Zavaterelli, M., Baretta, J.W., Baretta-Bekker, J.G., Pinardi, N.,

2000. The dynamics of the Adriatic Sea ecosystem. An idealized

model study. Deep-Sea Research I 47, 937–970.


planktonic community structure and carbon cycling in the ... · planktonic community structure and...

Documents